Greater Addition of Neurons to the Olfactory Bulb Than to the Cerebral Cortex of Eulipotyphlans but Not Rodents, Afrotherians Or Primates

ORIGINAL RESEARCH ARTICLE published: 11 April 2014 NEUROANATOMY doi: 10.3389/fnana.2014.00023 Greater addition of neurons to the olfactory bulb than to the cerebral cortex of eulipotyphlans but not rodents, afrotherians or primates

Pedro F. M. Ribeiro 1,2,PaulR.Manger3, Kenneth C. Catania 4, Jon H. Kaas 5 and Suzana Herculano-Houzel 1,2*

1 Instituto de Ciências Biomédicas, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brasil 2 Instituto Nacional de Neurociência Translacional, Ministério de Ciência e Tecnologia, São Paulo, Brasil 3 Department of Anatomy, University of the Witwatersrand, Johannesburg, South Africa 4 Department of Biology, Vanderbilt University, Nashville, TN, USA 5 Department of Psychology, Vanderbilt University, Nashville, TN, USA

Edited by: The olfactory bulb is an evolutionarily old structure that antedates the appearance of a Chet C. Sherwood, George six-layered mammalian cerebral cortex. As such, the neuronal scaling rules that apply to Washington University, USA scaling the mass of the olfactory bulb as a function of its number of neurons might be Reviewed by: shared across mammalian groups, as we have found to be the case for the ensemble Gordon M. Shepherd, Yale University School of Medicine, USA of non-cortical, non-cerebellar brain structures. Alternatively, the neuronal scaling rules Emmanuel Paul Gilissen, Royal that apply to the olfactory bulb might be distinct in those mammals that rely heavily Museum for Central Africa, Belgium on olfaction. The group previously referred to as Insectivora includes small mammals, *Correspondence: some of which are now placed in Afrotheria, a base group in mammalian radiation, Suzana Herculano-Houzel, Instituto and others in Eulipotyphla, a group derived later, at the base of Laurasiatheria. Here de Ciências Biomédicas, Universidade Federal do Rio de we show that the neuronal scaling rules that apply to building the olfactory bulb differ Janeiro, Av. Brigadeiro Trompowski, across eulipotyphlans and other mammals such that eulipotyphlans have more neurons s/n, Ilha do Fundão, Rio de Janeiro – concentrated in an olfactory bulb of similar size than afrotherians, glires and primates. RJ, 21941-590 Brazil Most strikingly, while the cerebral cortex gains neurons at a faster pace than the olfactory e-mail: [email protected] bulb in glires, and afrotherians follow this trend, it is the olfactory bulb that gains neurons at a faster pace than the cerebral cortex in eulipotyphlans, which contradicts the common view that the cerebral cortex is the fastest expanding structure in brain evolution. Our ﬁndings emphasize the importance of not using brain structure size as a proxy for numbers of neurons across mammalian orders, and are consistent with the notion that different selective pressures have acted upon the olfactory system of eulipotyphlans, glires and primates, with eulipotyphlans relying more on olfaction for their behavior than glires and primates. Surprisingly, however, the neuronal scaling rules for primates predict that the human olfactory bulb has as many neurons as the larger eulipotyphlan olfactory bulbs, which questions the classiﬁcation of humans as microsmatic.

Keywords: olfactory bulb, cortical expansion, mosaic evolution, olfaction

INTRODUCTION in the species examined—but not in Insectivora (Stephan and The olfactory bulb is an evolutionarily ancient structure of the Andy, 1969; Jolicoeur et al., 1984). This seemed consistent both forebrain, found in nearly all vertebrates, including jawless ﬁshes with a basal position of Insectivora and with the progressive view (Butler and Hodos, 2005). The olfactory bulb thus precedes the of evolution still in vogue in the late 20th century. appearance of a well-developed mammalian cerebral cortex, and At that time, Insectivora was a “scrap basket for small mam- its expansion drove the initial increase in relative brain size in mals” (Simpson, 1945),andviewedasanearlyoffshootin mammalian evolution (Rowe et al., 2011). However, the olfac- the placental mammal radiation, based on morphological cri- tory bulb has been shown to scale more slowly than the cerebral teria. In contrast, recent molecular data (Stanhope et al., 1998; cortex in a number of modern mammalian species, although Madsen et al., 2001; Murphy et al., 2001) provided convincing independently of the volume of the cerebral cortex across pri- evidence that the group was polyphyletic, and actually consisted mates (Stephan and Andy, 1969). While the relative volume of the of two orders: afrosoricida (tenrecs and golden moles), under cerebral cortex when regressed onto body mass was once shown to the superorder Afrotheria (Stanhope et al., 1998; Madsen et al., increase with brain volume across living mammalian species, and 2001; Murphy et al., 2001), and Eulipotyphla (hedgehogs, moles, to increase the most amongst primates (leading to the concept of shrews, solenodons; Waddell et al., 1999; Douady et al., 2002; cortical expansion in evolution; Stephan and Andy, 1969), the rel- Waddell and Shelley, 2003). Eulipotyphlans were later grouped ative volume of the olfactory bulb decreases by the same measure into the Laurasiatheria clade, which includes bats, cetartiodactyls,

Frontiers in Neuroanatomy www.frontiersin.org April 2014 | Volume 8 | Article 23 | 1 Ribeiro et al. Numbers of neurons in the olfactory bulb

carnivores, pangolins, and perissodactyls (Madsen et al., 2001; MATERIALS AND METHODS Murphy et al., 2001). Thus, in contrast from the original view that We applied the isotropic fractionator (Herculano-Houzel and placed insectivores as a whole as an early branch at the root of Lent, 2005) to determine the cellular composition of the olfac- placental mammals, this place is now occupied by Afrotherians tory bulbs of 10 glire species and 6 primate and the closely related (which do include golden moles, some of the original insecti- Scandentia species. Additionally, we compare our data to the cel- vores) and Xenarthra (Madsen et al., 2001; Murphy et al., 2001). lular composition of the olfactory bulbs of 5 insectivore species The remainder of the original insectivores, now known to be (Sarko et al., 2009) and 3 afrotherian species (Neves et al., 2014), eulipotyphlans, are considered a later offshoot, at the base of analyzed previously in the same manner. Laurasiatheria (Madsen et al., 2001; Murphy et al., 2001). We have previously demonstrated that the scaling rules that ANIMALS relate the mass of the cerebral cortex and cerebellum to their A total of 73 olfactory bulbs were collected from primates, rodents numbers of neurons differ across primates, glires and eulipoty- and one scandentia whose brains were prepared for the exami- phlans (reviewed in Herculano-Houzel, 2011). While in primates nation of cerebral structures in unrelated experiments. From pri- these structures scale linearly in mass with their number of mates, 25 olfactory bulbs were collected from 6 galagos (Otolemur neurons (Herculano-Houzel et al., 2007), in glires (rodents and Garnettii, 11 bulbs), 3 marmosets (Callithrix jacchus, 5 bulbs), lagomorphs) they grow in mass with the number of neurons 3owlmonkeys(Aotus trivirgatus, 6 bulbs), one rhesus monkey raised to the power of 1.7 and 1.2, respectively, which indicates (Macaca mulatta, 1 bulb), and one mouse lemur (Microcebus sp., that average neuronal size increases with number of neurons 2 bulbs). From Glires (rodents plus lagomorphs), we collected a (Herculano-Houzel et al., 2006, 2011). In eulipotyphlans, the total of 32 olfactory bulbs from 9 gray squirrels (Sciureus caroli- cerebral cortex scales in mass faster than it gains neurons, as it nensis, 9 bulbs), 5 Wistar rats (Rattus norvegicus, 5 bulbs), 2 agouti does in rodents, while the cerebellum increases in mass as a lin- (Dasyprocta primnolopha, 2 bulbs), 2 swiss mice (Mus musculus, ear function of its numbers of neurons, as in primates (Sarko 4 bulbs), 2 golden hamsters (Mesocrycetus auratus, 2 bulbs), 2 et al., 2009). These differences suggest that the cerebral cortex and guinea pigs (Cavia porcellus, 2 bulbs), 2 capybaras (Hydrochoerus cerebellum of primate, glires and eulipotyphlan brains are built hydrochoeris, 2 bulbs), 1 rabbit (Oryctolagus cuniculus, 1 bulb), according to cellular scaling rules that have diverged from their 3 naked mole-rats (Heterocephalus glaber,3bulbs)andone common ancestor. In contrast, the cellular scaling rules that apply Cayenne spiny rat (Proechimys guyannensis,2bulbs);someof to the remaining areas of the brain, excluding the olfactory bulb the present data on olfactory bulbs were published previously in (that is, the ensemble of brainstem, diencephalon and basal gan- Herculano-Houzel et al. (2011). We also collected 16 olfactory glia), appear to be more similar across these three mammalian bulbsfrom10treeshrews(Tupaia glis). All animals were adults orders, and might possibly be shared among them (Sarko et al., at the time of the experiments. All non-human primates, except 2009; Herculano-Houzel, 2011). This raises the possibility that for the mouse lemur, were raised in a colony in the Department of the olfactory bulb, being a structure as evolutionarily ancient as Psychology at Vanderbilt University. The mouse lemur brain was the brainstem and antedating the expansion of the cerebral cor- acquired from the Duke University Lemur Center. tex, might also vary in mass according to neuronal scaling rules Insectivore data were obtained from Sarko et al. (2009) and that are shared across mammalian orders. consist of the species averages of 34 olfactory bulbs from 5 species: Alternatively, the neuronal scaling rules that apply to the olfac- short-tailed shrew (Blarina carolinensis, 5 animals, 10 bulbs), star- tory bulb may have been preserved in some mammalian orders, nosed mole (Condylura cristata, 4 animals, 8 bulbs), smoky shrew but not in others, in a pattern that might shed light on the his- (Sorex fumeus, 3 animals, 6 bulbs), hairy-tailed mole (Parascalops tory of brain evolution. For instance, if afrotherians are an early breweri, 3 animals, 6 bulbs) and eastern mole (Scalopus aquaticus, diverging group in mammalian evolution, any scaling rules that 3 animals, 6 bulbs). Afrotherian data were obtained from Neves are shared by this group and others might be a characteristic of the et al. (2014) and consist of species averages of 5 olfactory bulbs common ancestor to eutherians. Here we investigate the rules that from 3 species: elephant shrew (Elephantulus myurus, 2 animals, apply to the scaling of the mass of the olfactory bulb as a function 2 bulbs), four-toed elephant shrew (Petrodomus tetradactylus,2 of its number of neurons in glires, primates, and the two clades animals, 2 bulbs), and rock hyrax (Procavia capensis, 1 animal, 1 once lumped as Insectivora: afrotherians and eulipotyphlans, by bulb). using the isotropic fractionator (Herculano-Houzel and Lent, 2005) to determine the cellular composition of the olfactory bulb TISSUE in 24 species, a method that was recently shown to give results Olfactory bulbs were processed separately for each hemisphere; in that are comparable to those obtained with stereology (Bahney most cases, both olfactory bulbs were available. Since no signiﬁ- and von Bartheld, 2014). Further, we investigate how numbers cant differences were observed between left and right olfactory of neurons in the olfactory bulb relate to numbers of neurons in bulb mass or number of neurons (2-tailed Student t-test, p > the whole brain (without the olfactory bulb) and particularly in 0.05), all data were pooled together and presented as average the cerebral cortex across species, to examine the notion that the values for both left and right olfactory bulbs for each species. cerebral cortex is the fastest-scaling structure in the brain and its Numbers of neurons in other brain structures of the respec- corollary: that the decrease in relative size of the olfactory bulb is tive species were obtained from our previous studies using the equivalent to a decrease in the relative number of neurons in the same counting methodology (Herculano-Houzel et al., 2006, olfactory bulb as the brain gains neurons across species. 2007; Sarko et al., 2009; Gabi et al., 2010; Neves et al., 2014). All

Frontiers in Neuroanatomy www.frontiersin.org April 2014 | Volume 8 | Article 23 | 2 Ribeiro et al. Numbers of neurons in the olfactory bulb

numbers presented here refer to both olfactory bulbs on the two at present quantify them. We focused, instead, on total numbers sides of the brain, and to the whole brain, cerebral cortex or cere- of neurons in the olfactory bulb, which are mostly located in bellum, as described in our previous studies (Herculano-Houzel the granule cell layer and in the glomerular layer. In the mouse et al., 2006, 2007, 2011; Sarko et al., 2009; Gabi et al., 2010). main olfactory bulb, for instance, 52.6 and 34.0% of all neurons are found in the glomerular and granule cell layers, respectively DISSECTION (Parrish-Aungst et al., 2007). Given that non-neuronal cells are All primates and scandentia, except for the mouse lemur, were typically at least half of all cells in the olfactory bulb, we consider sacrificed by lethal injection of sodium pentobarbital, and per- that the misplacement of mitral cell neurons as non-neuronal fused transcardially with 0.9% phosphate-buffered saline fol- cells does not significantly influence the estimates of total num- lowed by 4% phosphate-buffered paraformaldehyde. Rodents bers of neurons or non-neuronal cells in the main olfactory bulb were sacrificed by inhalation of ether and perfused transcardially reported here. as above. Brains were removed from the skull with care not to damage the olfactory bulbs, which were then removed by tran- DATA ANALYSIS secting the olfactory tract immediately proximal to the bulb. Due All statistical analyses and regressions were performed in JMP 9.0 to variation across species, this procedure may include unknown (SAS, USA), using the average values obtained from the individu- amounts of the anterior olfactory nucleus in the olfactory bulb. als of each species. Correlations between variables were calculated However, for the sake of consistency, all care was taken to ensure using the Spearman correlation coefficient. If a significance cri- that the transaction of the olfactory tract was performed at a terion of p < 0.05 was reached, regressions of the data to linear level where the olfactory tract was completely exposed, leaving all and power functions were calculated. Exponents are presented ± tissue surrounded by gray matter included in the structure that standard error. In all analyses, “whole brain” refers to the sum we refer to as the olfactory bulb. The olfactory bulbs were then of cerebral cortex, cerebellum and remaining areas, not including weighed, and post-fixed for 2 weeks to 12 months by immersion the olfactory bulb, as in our previous studies (Herculano-Houzel in 4% phosphate-buffered paraformaldehyde. et al., 2007, 2011; Sarko et al., 2009; Gabi et al., 2010). Analysis of numbers of cells exclude the capybara, for which the quality of ISOTROPIC FRACTIONATOR the preparation was doubtful, yielding numbers of cells that were Total numbers of cells, neurons, and non-neuronal cells were esti- less than half the expected. mated as described previously using the Isotropic Fractionator method (Herculano-Houzel and Lent, 2005). Briefly, each olfac- RESULTS tory bulb is turned into an isotropic suspension of isolated nuclei Olfactory bulb size varies greatly within and between mammalian of known, defined volume, kept homogeneous by agitation. The orders. In our sample of 24 species of glires, eulipotyphlans, total number of nuclei in suspension—and therefore the total afrotherians, primates, and scandentia, we find a 162.8-fold vari- number of cells in the original tissue—is estimated by deter- ation between the smallest and the largest olfactory bulb (0.008 g mining the density of nuclei in small aliquots stained with the in the marmoset against 1.302 g in the capybara). Within orders, fluorescent DNA marker DAPI (4-6-diamidino-2-phenylindole olfactory bulb size varies 25-fold in primates (from 0.008 g in the dihydrochloride; Invitrogen), under the microscope. Once the marmoset to 0.200 g in the galago), 93-fold in glires (from 0.014 g total cell number is known, the proportion of neurons is deter- in the mouse to 1.302 g in the capybara), and 7x in eulipotyphlans mined by immunocytochemical detection of Neuronal Nuclear (from 0.012 g in the smoky shrew to 0.082 g in the eastern mole; antigen (NeuN, mab377, Millipore), expressed in all nuclei of Table 1). The range of variation in olfactory bulb size is however most neuronal cell types and not in non-neuronal cells (Mullen overlapping across the orders, and 13 out of 21 species have olfac- et al., 1992). Estimates of the proportion of NeuN-positive nuclei tory bulbs within the range of 0.01–0.1 g (Figure 1A). The two are considered reliable since the coefficient of variation among largest olfactory bulbs belong to rodents (agouti and capybara), animals of the same species is typically well below 0.15. Numbers which are remarkably large compared to the other bulbs analyzed. of non-neuronal cells are derived by subtraction. Mitral cells in the olfactory bulb constitute the main output of OLFACTORY BULB MASS the olfactory bulb, along with tufted cells, and are one of the neu- Across species, larger glire and insectivore brains (excluding the ronal subtypes known not to express NeuN (Mullen et al., 1992). olfactory bulb) tend to have larger olfactory bulbs, with over- Mitral cell neurons are therefore included in the NeuN-negative lapping distributions of brain and olfactory bulb mass that (other, non-neuronal cells) population in our study, and thus arti- also include afrotherians (Figure 1A). Olfactory bulb mass in ficially inflate the count of non-neuronal cells in the olfactory insectivores varies as a power function of brain mass (which bulb, while tufted cells are supposedly included as NeuN-positive doesn’t include the olfactory bulb) with exponent 0.965 ± neurons. However, mitral cells are vastly outnumbered by granule 0.149 (r2 = 0.933, p = 0.0075), which is indistinguishable from cell neurons and other cells in the olfactory bulb. In the mouse, unity (Figure 1A). Indeed, the relative mass of the eulipoty- the mitral cell layer (which includes other, GABAergic cell types) phlan olfactory bulb in proportion to the brain does not vary represents only 7.9% of all neurons in the olfactory bulb of the systematically with increasing brain mass (Spearmann correla- mouse, and tufted cells are estimated at 3–6% of all neurons tion, ρ = 0.100, p = 0.8729; Figure 1B), and ranges between (Parrish-Aungst et al., 2007). While it will definitely be impor- 4.99 and 8.21%. In glires, olfactory bulb mass varies as a tant to quantify mitral and tufted cells separately, we could not power function of brain mass of exponent 0.820 ± 0.074 (p <

Frontiers in Neuroanatomy www.frontiersin.org April 2014 | Volume 8 | Article 23 | 3 Ribeiro et al. Numbers of neurons in the olfactory bulb

Table 1 | Cellular composition of the olfactory bulb.

Species Mass, g Neurons % Neurons Neurons/mg

GLIRES Mus musculus 0.014 ± 0.004 3.89 × 106 ± 1. 2 5 ± 106 41.1 ± 4.3 257,475 ± 34,036 Mesocricetus auratus 0.055 ± 0.011 5.75 × 106 ± 0.35 ± 106 52.1 ± 12.0 105,418 ± 27,498 Rattus norvegicus 0.074 ± 0.022 11.10 × 106 ± 3.20 ± 106 54.4 ± 4.1 152,373 ± 26,913 Proechimys cayennensis 0.132 9.14 × 106 30.2 69,254 Cavia porcellus 0.103 ± 0.013 6.06 × 106 ± 1. 3 0 ± 106 38.2 ± 5.0 58,560 ± 5,340 Sciurus carolinensis 0.212 ± 0.022 28.84 × 106 ± 7. 9 0 ± 106 43.4 ± 12.1 137,532 ± 38,236 Oryctolagus cuniculus 0.156 18.76 × 106 45.0 120,288 Dasyprocta primnolopha 0.737 ± 0.162 58.12 × 106 ± 4.95 ± 106 44.9 ± 4.7 88,008 ± 14,973 Hydrochoerus hydrochoeris 1.302 ± 0.031 28.56 × 106 ± 8.52 ± 106 30.0 ± 1.6 21,864 ± 6018 PRIMATES Microcebus murinus 0.030 ± 0.008 7.64 × 106 ± 0.12 ± 106 44.0 ± 1.8 270,894 ± 61,946 Callithrix jacchus 0.008 ± 0.014 2.11 × 106 ± 0.98 ± 106 45.2 ± 7.2 232,309 ± 137,605 Otolemur garnetti 0.200 ± 0.016 30.24 × 106 ± 9.64 ± 106 46.3 ± 7.3 149,219 ± 43,590 Aotus trivirgatus 0.050 ± 0.012 7.92 × 106 ± 3.11 ± 106 48.4 ± 11.5 155,879 ± 62,241 Macaca mulatta 0.088 8.47 × 106 43.5 96,293 SCANDENTIA Tupaia glis 0.100 ± 0.032 12.7 × 106 ± 3.58 ± 106 38.3 ± 6.0 130,173 ± 17,451 EULIPOTYPHLA Sorex fumeus 0.012 ± 0.002 3.33 × 106 ± 1. 0 5 ± 106 53.8 ± 6.4 289,806 ± 124,350 Blarina brevicauda 0.026 ± 0.003 8.09 × 106 ± 0.94 ± 106 62.2 ± 4.5 318,164 ± 34,950 Parascalops breweri 0.049 ± 0.008 16.75 × 106 ± 6.37 ± 106 60.5 ± 5.6 333,590 ± 81,590 Condylura cristata 0.040 ± 0.005 10.55 × 106 ± 4.29 ± 106 57.1 ± 9.2 254,720 ± 74,620 Scalopus aquaticus 0.082 ± 0.005 34.61 × 106 ± 5.96 ± 106 65.8 ± 6.5 423,520 ± 94,950 AFROTHERIA Petrodomus tetradactylus 0.159 ± 0.013 12.83 × 106 ± 0.54 ± 106 48.2 ± 11.8 80,805 ± 3084 Elephantulus myurus 0.050 ± 0.014 9.69 × 106 ± 2.47 ± 106 66.5 ± 1.5 194,678 ± 5708 Procavia capensis 0.286 20.91 × 106 58.6 73,110

All values refer to both olfactory bulbs. % Neurons, percentage of cells in the olfactory bulb that are NeuN+. Neurons/mg, density of neurons per miligram of olfactory bulb tissue. Data from Sarko et al. (2009), Herculano-Houzel et al. (2011), Neves et al. (2014) and this study.

0.0001), which is significantly different from unity. As a con- The mass of the olfactory bulb varies linearly with the mass of sequence, the relative mass of the olfactory bulbs in propor- the cerebral cortex across eulipotyphlans (power exponent, 0.930 tion to the brain (average of 3.94 ± 0.50%) decreases with ± 0.170, p = 0.0120), and slightly below linearity across glires increasing brain mass in glires (exponent −0.180 ± 0.0074, (power exponent, 0.772 ± 0.073, p < 0.0001). Afrotherians over- p = 0.0409), varying from 1.71% (in the rabbit) to 6.35% of lap with both eulipotyphlans and glires. There is no significant brain mass (in the spiny rat; Figure 1B). Afrotherians overlap correlation between olfactory bulb mass and cortical mass across with both glires, with an average relative mass of the olfac- primates (Figure 1C). tory bulbs of 4.34 ± 1.41% of brain mass. Across primate and scandentia species, in turn, there is no significant correlation CELLULAR SCALING RULES FOR THE OLFACTORY BULB between olfactory bulb mass and brain mass (Spearmann cor- Despite the different absolute and relative sizes of the olfac- relation, ρ = 0.429, p = 0.3374; Figure 1A). Thus, while larger tory bulb across species and orders, the relationship between brains are accompanied by larger olfactory bulbs across both olfactory bulb mass and its number of non-neuronal cells is over- eulipotyphlan and glire species, the relative size of the olfac- lapping among eulipotyphlans, glires, afrotherians and primates tory bulbs compared to the brain does neither increase nor (Figure 2A). For the ensemble of species, olfactory bulb mass decrease with brain size in insectivores or primates, but it can be expressed equally well as a linear function of its numbers decreases with increasing brain size in glires. However, there is of non-neuronal cells (r2 = 0.847, p < 0.0001) and as a power a significant negative correlation between the relative mass of function of the number of these cells with exponent 1.164 ± the olfactory bulb and brain mass across all orders combined 0.115, which includes unity (p < 0.0001; Figure 2A). This indi- (Spearman correlation, ρ =−0.736, p < 0.0001; Figure 1B). The cates that the olfactory bulbs of eulipotyphlans, glires, afrotheri- relative mass of the olfactory bulb is smaller in primates (0.01– ans, primates, and scandentia share non-neuronal scaling rules, 3.63% of brain mass) than in glires, and smaller in glires which agrees with previous observations for the cerebral cortex, than in eulipotyphlans, for a comparable range of brain mass cerebellum and remaining brain areas across mammalian orders (Figure 1B). (Herculano-Houzel, 2011).

Frontiers in Neuroanatomy www.frontiersin.org April 2014 | Volume 8 | Article 23 | 4 Ribeiro et al. Numbers of neurons in the olfactory bulb

FIGURE 1 | Continued 0.170, p = 0.0120) and sublinearly across glires (green; exponent, 0.772 ± 0.073, p < 0.0001), but there is no signiﬁcant correlation between olfactory bulb mass and cortical mass across primates (red). Eulipotyphlan data from Sarko et al. (2009). Afrotherian data from Neves et al. (2014). Glires, data from Herculano-Houzel et al. (2011) and this study. Primates and scandentia, data from this study.

FIGURE 2 | Neuronal and non-neuronal scaling rules for the olfactory bulb. Graphs show the relationship between average mass of both olfactory bulbs in each species and (A) the total number of other FIGURE 1 | Distribution of absolute and relative olfactory bulb mass (non-neuronal) cells or (B) the total number of neurons in them. (A) across species. (A) Average mass of both olfactory bulbs in each species Olfactory bulb mass varies similarly, and linearly, with its number of other and (B) relative mass of the olfactory bulbs expressed as a percentage of cells across the ensemble of eulipotyphlans (black), glires (green), primates, brain mass, both shown as a function of absolute brain mass (which does and scandentia (red and orange), but (B) as a power function of its number not include the olfactory bulbs). Olfactory bulb mass varies linearly with of neurons with an exponent of 0.823 ± 0.071 in eulipotyphlans, 1.092 ± brain mass in eulipotyphlans (A, black; power function exponent, 0.965 ± 0.170 in glires, and 1.243 ± 0.192 in primates and scandentia. Notice that 0.149, p = 0.0075), such that its relative size does not vary systematically eulipotyphlans have more neurons packed into olfactory bulbs of a similar with brain mass (B, black). In glires, olfactory bulb mass increases more mass in glires or afrotherians. Eulipotyphlan data from Sarko et al. (2009). slowly than brain mass (A, green; power function exponent, 0.820 ± 0.074, Afrotherian data from Neves et al. (2014). Glires, data from p < 0.0001), such that the relative mass of the olfactory bulb decreases Herculano-Houzel et al. (2011) and this study. Primates and scandentia, data with increasing brain mass (B, green). Afrotherians (blue) overlap with both from this study. eulipotyphlans and glires. In primates and scandentia there is no signiﬁcant correlation between olfactory bulb mass and brain mass (A and B,redand orange). (C) The mass of the olfactory bulb increases linearly with the In contrast, different sets of neuronal scaling rules apply to the ± mass of the cerebral cortex across eulipotyphlans (black; exponent, 0.930 olfactory bulb of primates, glires and eulipotyphlans. Variations (Continued) in olfactory bulb mass are best described as a linear function

Frontiers in Neuroanatomy www.frontiersin.org April 2014 | Volume 8 | Article 23 | 5 Ribeiro et al. Numbers of neurons in the olfactory bulb

ofthenumberofneuronsinthestructurebothinprimates (r2 = 0.929, p = 0.0020, against power exponent 1.243 ± 0.192, r2 = 0.912, p = 0.0030) and in glires (r2 = 0.937, p = 0.0002, against power exponent r2 = 0.856, 1.092 ± 0.170, p = 0.0004; Figure 2B.). Across eulipotyphlans, however, olfactory bulb mass is best described as a power law of the number of neurons with a smaller exponent of 0.823 ± 0.071, which excludes linearity (r2 = 0.978, p < 0.0014; Figure 2B). Afrotherians overlap with therelationshipforglires,noteulipotyphlans(Figure 2B,blue). Combined with the shared, near-linear scaling of olfactory bulb mass with numbers of non-neuronal cells, these exponents suggest that average neuronal size in the olfactory bulb remains constant with bulb size across primate and glires species, but decreases as olfactory bulb mass increases in eulipotyphlans (Sarko et al., 2009). The direct comparison across species shows that olfactory bulbs of a similar mass above about 0.02 g are built with larger numbers of neurons in insectivores than in primates, glires or afrotherians—that is, similar numbers of neurons are packed into smaller olfactory bulbs in eulipotyphlans compared to primates, glires and afrotherians (Figure 2B). In agreement with the shared and linear non-neuronal scaling rules among orders, we find overlapping non-neuronal cell densities across most of the species that do not vary in correlation with bulb mass (Spearman correlation, all p > 0.3; Figure 3A). In contrast, eulipotyphlans exhibit much larger neuronal densities in the olfactory bulb than afrotherians, most glires, or primates (Figure 3B). In line with a linear relationship between bulb mass and number of neurons across glires and primates, we find no significant correlation between neuronal density and olfactory bulb mass in either group (Spearman correlation, p > 0.1; Figure 3B). In eulipotyphlans, as described before (Sarko et al., 2009), neuronal density does not vary significantly with bulb mass across the five species (Spearman correlation, ρ = 0.700, p = 0.1881). However, the star-nosed mole is an outlier in the group, with a FIGURE 3 | Neuronal and non-neuronal cell densities in the olfactory bulb. Graphs show the relationship between olfactory bulb mass and (A) much smaller neuronal density than would be expected based on average non-neuronal (other) cell density or (B) average neuronal density in the other four insectivore species examined (Sarko et al., 2009). If the olfactory bulb across species. (A) There is no significant correlation the star-nosed mole species is removed from the comparison, we between other cell density (in other cells/mg) and olfactory bulb mass find that neuronal density in the olfactory bulb increases signifi- across species within any order. (B) Neuronal cell density does not ρ = . =< . correlate with olfactory bulb mass in primates (and scandentia) or glires, cantly with bulb mass ( 1 000, p 0 0001). As mentioned but it does increase with larger olfactory bulb mass in eulipotyphlans when above, these data suggest that average neuronal size in the olfac- the outlier, the star-nosed mole, is excluded. Notice that neuronal densities tory bulb does not vary systematically with increasing bulb size in in the olfactory bulb in eulipotyphlan species are higher than in most other primates of glires, but decreases in eulipotyphlans. species. Eulipotyphlan data from Sarko et al. (2009). Afrotherian data from Consistently with the larger neuronal densities in insecti- Neves et al. (2014). Glires, data from Herculano-Houzel et al. (2011) and this study. Primates and scandentia, data from this study. vores than in the other two groups, eulipotyphlans also differ from the other orders in the percentage of olfactory bulb cells that are neurons. This percentage is higher in eulipotyphlans ± ± (59.9 2.1%) than in afrotherians (48.0 10.8%), primates p = 0.0374, 0.0005, and 0.6228, respectively; Figure 4A). In both ± ± and scandentia (44.4 1.3%) or glires (43.2 2.4%; ANOVA, eulipotyphlans and glires, the rate of addition of neurons to the = . p 0 0101). olfactory bulb as a function of number of neurons in the whole brain is indistinguishable from linearity (power exponents, 1.163 RELATIVE DISTRIBUTION OF NEURONS IN THE OLFACTORY BULB ± 0.247 and 0.884 ± 0.122, r2 = 0.881 and 0.883, p = 0.0182 and COMPARED TO THE WHOLE BRAIN 0.0002, respectively; linear fit, r2 = 0.856 and 0.896, p = 0.0243 Just as larger glire and eulipotyphlan (but not primate) brains and 0.0001; Figure 4A). accompany larger olfactory bulbs (Figure 1), larger numbers of Eulipotyphlans,however,havemoreneuronsintheolfactory neurons in the brain are accompanied by larger numbers of neu- bulb than glires and afrotherians for a similar number of brain rons in the olfactory bulb in eulipotyphlans and glires, but not neurons (Figure 4A). Indeed, the eulipotyphlan olfactory bulb in primates (Spearman correlation, ρ = 0.900, 0.891, and 0.257, has on average 12.5 ± 1.7% of the total number of neurons

Frontiers in Neuroanatomy www.frontiersin.org April 2014 | Volume 8 | Article 23 | 6 Ribeiro et al. Numbers of neurons in the olfactory bulb

FIGURE 4 | Scaling of numbers of neurons in the olfactory bulb as a function of numbers of neurons in the brain and in the cerebral cortex. (A) The number of neurons in the olfactory bulb increases as power functions of the number of neurons in the whole brain (without the olfactory bulb) of exponent 1.163 ± 0.247 in eulipotyphlans and 0.884 ± 0.122 in glires, respectively, but not in primates. Notice that eulipotyphlans have more neurons in the olfactory bulb than glires and afrotherians for similar numbers of brain neurons. (B) The number of neurons in the olfactory bulb of eulipotyphlans increases at a faster rate than in the cerebral cortex, as a power function of exponent 2.129 ± 0.428, but at the same pace as the cerebral cortex in glires, with an exponent of 1.018 ± 0.148. Notice that eulipotyphlans have more neurons in the olfactory bulb than glires and afrotherians for similar numbers of cortical neurons. There is FIGURE 5 | Scaling of relative numbers of neurons in the olfactory bulb no significant relationship between numbers of olfactory bulb and cerebral compared to the brain and cerebral cortex. (A) The relative number of cortical neurons across primate species. Eulipotyphlan data from Sarko neurons in the olfactory bulb expressed as a percentage of the number of et al. (2009). Afrotherian data from Neves et al. (2014). Glires, data from neurons in the brain (excluding the olfactory bulb) is not significantly Herculano-Houzel et al. (2011) and this study. Primates and scandentia, data correlated with brain mass within any of the mammalian orders examined, from this study. although it decreases with increasing brain mass across all species combined. (B) The relative number of neurons in the olfactory bulb in the brain (which does not include the olfactory bulb), while increases with increasing relative mass of the olfactory bulb (both ± compared to the whole brain) across primates, glires and afrotherians in afrotherians the olfactory bulb has 6.3 1.5% of the neu- ± ± combined, as a power function of exponent 0.877 0.054. (C) The relative rons in the brain, in glires it has 5.3 0.7% of the number number of neurons in the olfactory bulb expressed as a percentage of the found in the brain, and in primates, only 2.0 ± 0.8% (ANOVA, number of neurons in the cerebral cortex is not significantly correlated with p < 0.0001; Figure 5A). As expected from the linear relationships brain mass within any of the mammalian orders examined. Eulipotyphlan between numbers of neurons in the olfactory bulb and in the data from Sarko et al. (2009). Afrotherian data from Neves et al. (2014). brain, in none of the orders is the relative number of neurons Glires, data from Herculano-Houzel et al. (2011) and this study. Primates and scandentia, data from this study. in the olfactory bulb expressed as a percentage of the number of

Frontiers in Neuroanatomy www.frontiersin.org April 2014 | Volume 8 | Article 23 | 7 Ribeiro et al. Numbers of neurons in the olfactory bulb

brain neurons significantly correlated with brain mass (Spearman correlations: insectivores, p = 0.6238; glires, 0.1739; primates and scandentia, 0.2080; Figure 5A). Thus, numbers of neurons increase concertedly between the olfactory bulb and the brain in eulipotyphlans and glires (while the former have larger absolute and relative numbers of neurons in the olfactory bulb than the latter for a same number of brain neurons), but are independent of one another in primates. However, there is a significant negative correlation between the relative number of neurons in the olfactory bulb and brain mass across all orders (Spearman correlation, ρ =−0.728, p < 0.0001; Figure 5A). Because the relative mass of the olfactory bulb is also negatively correlated with brain mass across all orders (Figure 1B), the relationship between relative number of neurons and relative mass of the olfactory bulb (both expressed as percentages of whole brain) can be described as a single power function across primates, glires and scandentia, with an exponent that approaches linearity (0.877 ± 0.054, r2 = 0.942, p < 0.0001), although 3 species of Eulipotyphla remain outliers in the relationship (Figure 5B). Although the eulipotyphlan olfactory bulb gains neurons linearly with the brain as a whole, it gains neurons at a faster rate than the cerebral cortex, as a power function of the number of cortical neurons of exponent 2.129 ± 0.428, which excludes unity and, therefore, linearity (Figure 4B). In eulipotyphlans, the number of neurons in the olfactory bulb represents on average 75.0 ± 16.5%ofthenumberofneuronsinthecerebralcortex,andactu- ally exceeds the latter in two species (eastern mole and hairy-tailed mole, the two with the largest olfactory bulbs amongst insectivores in the sample; Figure 5C). As expected from the supralinear rate of addition of neurons to the olfactory bulb compared to the cerebral cortex, the relative number of neurons in the former structure expressed as a percentage of the latter increases significantly with brain mass across eulipotyphlans (Spearman correlation, ρ = 0.900, p = 0.0374). In constrast, the olfactory FIGURE 6 | Scaling of numbers of neurons in the olfactory bulb and in bulb in glires gains neurons at the same pace as the cerebral cortex the cerebral cortex as a function of numbers of neurons in the rest of (power exponent, 1.018 ± 0.148, r2 = 0.872, p = 0.0002; linear brain. (A) The number of neurons in the olfactory bulb increases as power 2 = . = . functions of the number of neurons in the rest of brain of exponent 1.770 ± fit, r 0 862, p 0 0003; Figure 4B), and represents on aver- ± ± 0.578 in eulipotyphlans and 1.198 0.252 in glires, respectively, but not in age only 31.0 3.9% of the neurons found in the cerebral cortex, primates. Notice that eulipotyphlans have more neurons in the olfactory bulb a percentage that does not vary significantly together with brain than glires and afrotherians for similar numbers of rest of brain neurons. (B) mass in glires (Spearman correlation, p = 0.4888). Similarly, the The number of neurons in the cerebral cortex of eulipotyphlans increases afrotherians examined have a number of neurons in the olfac- more slowly than the number of neurons in the rest of brain, as a power ± function of exponent 0.846 ± 0.180, but at the same pace as the cerebral tory bulb that represents 28.6 9.0% of the number of neurons cortex in glires, with an exponent of 1.018 ± 0.148. Notice that eulipotyphlans found in the cerebral cortex (Figure 5C). Eulipotyphlans, in com- have more neurons in the olfactory bulb than glires and afrotherians for similar parison, have more neurons in the olfactory bulb than glires numbers of cortical neurons. There is no significant relationship between and afrotherians for a similar number of neurons in the cerebral numbers of olfactory bulb and cerebral cortical neurons across primate cortex (Figure 4B). In contrast, in primates and scandentia, the species. Eulipotyphlan data from Sarko et al. (2009). Afrotherian data from ± Neves et al. (2014). Glires, data from Herculano-Houzel et al. (2011) and this number of neurons in the olfactory bulb represents only 12.0 study. Primates and scandentia, data from this study. 5.6% of the number of neurons in the cerebral cortex, and there is no significant correlation between numbers of neurons across the two structures (Spearman correlation, p = 0.2080; Figure 4B). in the rest of brain (Figure 6A). In eulipotyphlans, the cerebral The eulipotyphlan olfactory bulb gains neurons at a faster rate cortex gains neurons more slowly than the rest of brain (power than the non-cortical, non-cerebellar areas of the brain (“rest exponent, 0.846 ± 0.180, r2 = 0.880, p = 0.0182; Figure 6B), of brain”), as a power function of the number of rest of brain although this relationship can also be described as a linear func- neurons of exponent 1.770 ± 0.578 (r2 = 0.758, p = 0.0548; tion (r2 = 0.936, p = 0.0069). Despite the statistical uncertainty, Figure 6A) or as a linear function (r2 = 0.906, p = 0.0127). The the most likely scenario that best accounts for the scaling of olfactory bulb also has more neurons in eulipotyphlans than in numbers of neurons in the olfactory bulb in eulipotyphlans is glires, afrotherians or primates with a similar number of neurons that it gains neurons faster than the rest of the brain, while both

Frontiers in Neuroanatomy www.frontiersin.org April 2014 | Volume 8 | Article 23 | 8 Ribeiro et al. Numbers of neurons in the olfactory bulb

structures gain neurons faster than the cerebral cortex, such that Primates of all brain sizes do have many more neurons in the cere- the olfactory bulb gains neurons even faster compared to the cor- bral cortex than in the olfactory bulb, which is consistent with tex than to the rest of brain. In glires, the olfactory bulb gains the expansion of other sensory systems, and eulipotyphlans have neurons at the same pace as the rest of brain (power exponent, as many or more neurons in the olfactory bulb than in the cere- 1.198 ± 0.252, which includes unity; r2 = 0.763, p = 0.0021; bral cortex, which is consistent with a heavy reliance on olfaction. linear fit, r2 = 0.816, p = 0.0008; Figure 6A); and in primates However, both groups have a similar range of numbers of neurons and scandentia, the number of neurons in the olfactory bulb in the olfactory bulb. We note that the distribution of projection is not significantly correlated with numbers of neurons in the and intrinsic neurons, both excitatory and inhibitory, is differ- restofbrain(powerfunction,p = 0.7902; Figure 6A). In con- ent across structures such as the cerebral cortex, the olfactory trast, the cerebral cortex gains neurons faster than the rest of bulb and the cerebellum, with presumably decreasing proportions brain in both glires and primates (exponents of 1.190 ± 0.110 of projection neurons across these three structures, respectively. and 1.723 ± 0.314, and p < 0.0001 and p = 0.0054, respectively; However, given that both projection and intrinsic neurons are the Figure 6B). The scaling relationships that apply across structures basic units of information processing, combining the activity of are summarized in Figure 7. thousands of synapses, we assume that total numbers of neurons are a valid proxy for total information processing capacity in each DISCUSSION structure. Thus, if total numbers of neurons in the olfactory bulb Here we find that the eulipotyphlan olfactory bulb gains neurons can be considered an indication of the amount of olfactory pro- faster than the cerebral cortex across species, an unexpected find- cessing that can be performed, then it must be conceded that the ing in face of the decrease in the relative mass of the olfactory bulb increasing reliance on other sensory systems such as vision in pri- and relative expansion of the cerebral cortex with increasing brain mates did not occur at the expense of olfaction, but in addition size. Moreover, in some eulipotyphlan species the olfactory bulb to it. has as many or more neurons than the cerebral cortex, despite Second, our finding that the neuronal scaling rules that apply the larger size of the latter. These findings have several implica- to the olfactory bulb appear to be shared by afrotherians and tions for mammalian brain evolution. First, the large and growing glires, but not eulipotyphlans or primates, concurs with the phy- absolute numbers of neurons in the olfactory bulb of eulipoty- logenetic analysis that shows that modern eulipotyphlans are a phlans compared to the cerebral cortex dispute the notion that, more recent group than afrotherians and actually do not reflect while olfaction was emphasized in early mammals, other sensory the ancestral mammalian state. A parsimonious alternative is that systems necessarily became more important as cortex expanded modern glires and afrotherians (and possibly primates as well) in evolution (Rowe et al., 2011), such as vision in primates (Gilad retained the ancestral neuronal rules for building the olfactory et al., 2004). The fact that eulipotyphlans are now considered a bulb and for adding neurons faster to the cerebral cortex than later branch than afrotherians in mammalian evolution argues for to the olfactory bulb in glires and afrotherians, while eulipoty- a renewed importance of olfaction in the behavior of this group. phlans evolved a new way of scaling the olfactory bulb, both with

FIGURE 7 | Scaling relationships for neurons and mass across larger than 1 is the larger side of the polygon and has its structures in each mammalian order. Lines indicate linear scaling in exponent indicated in the ﬁgure. Notice that the faster scaling of numbers of neurons (top) or mass (bottom) across structures; number of neurons in the eulipotyphlan olfactory bulb than in the polygons represent reciprocal power laws in the scaling of one cerebral cortex would not be predicted from the mass relationships structure against another, where the power law with the exponent across the two structures.

Frontiers in Neuroanatomy www.frontiersin.org April 2014 | Volume 8 | Article 23 | 9 Ribeiro et al. Numbers of neurons in the olfactory bulb

a higher density of neurons that allows for a larger number of neu- important now to look at the neuronal scaling rules that apply rons to be concentrated in a small volume, and with a higher rate to the olfactory bulb of other branches of placental mammals, of addition of neurons to the olfactory bulb than to the cerebral and to non-eutherian mammals: marsupials and monotremes. cortex. The prediction based on the current findings and hypothesis The olfactory system is known to be malleable in evolu- is that the cellular composition of the olfactory bulb of these tion, depending on use. Activation of odorant receptor genes latter groups should match those found for afrotherians and induces growth of the olfactory epithelium, which in turn induces glires. turbinal growth and ossification (Rowe et al., 2011), and might Here we show that, across species, the olfactory bulb gains neu- be related to the support of larger numbers of neurons in the rons faster than the cerebral cortex in insectivores; at the same rate olfactory bulb and pyriform cortex. The olfactory bulb also gains in glires; and independently in primates. This is in contrast to the neurons in adult life through the rostral migratory stream (Lois olfactory bulb gaining mass at the same rate as the cerebral cortex et al., 1996), and it is currently unknown whether an increased across insectivore species and more slowly than the cerebral cor- addition of neurons in adult life might contribute to the faster tex across glires. Our finding that the relative scaling of numbers scaling of neurons in the olfactory bulb than in the cerebral cortex of neurons in the olfactory bulb compared to other brain struc- of eulipotyphlans. Odontocete cetaceans lack an olfactory bulb tures differs from the scaling expected from mass relationships (Turner, 1891; Jacobs et al., 1971), and the structure is relatively underscores the importance of not using volume indiscriminately very small in other aquatic mammals such as pinnipeds and the as a proxy for numbers of neurons, which was common practice manatee (Reep et al., 2007), in line with the little use of chemi- in the literature (for instance, Finlay and Darlington, 1995; Clark cal olfaction in the water by these particular mammalian species. et al., 2001; Sultan, 2002). Similarly, structures related to olfaction exhibit mosaic evolution One of the original uses of brain structure volumes to make in other groups, changing in size coordinately amongst them- inferences about mammalian brain evolution was by Stephan selves but in a manner that is unrelated to other brain structures and Andy (1969), who put forward the notion that brain evolu- (Barton and Harvey, 2000). The greater addition of neurons to tion equates with cortical expansion in detriment of other brain the olfactory bulb than to the cerebral cortex in eulipotyphlans, in structures, and most of all olfactory bulb. Those authors used contrast to the opposite pattern in glires and the lack of a correla- progression indices (akin to encephalization, calculated for each tion in primates, implies that olfaction is an extremely important structure from the regression of their volume onto body mass) senseforeulipotyphlans,moresothanforothermammals.This to show that neocortex is the structure that becomes relatively conclusion is further strengthened by recent behavioral studies. increased the most, while olfactory bulb is the only structure to The eastern mole (Scalopus aquaticus) can navigate rapidly up “involute.” Using Stephan’s data, Finlay later repeated this analysis odor gradients without using scent trails to localize prey. This is and concluded that the olfactory bulb increases in volume faster accomplished with serial sampling over time (sniffing and mov- than total brain size in insectivores and bats, but more slowly than ing) combined with comparison of stereo nostril cues during each total brain size in simians and prosimians (Finlay et al., 2001). sniff (Catania, 2013). Both water shrews (Sorex palustris)and These conclusions contributed to the notion that, in mam- star-nosed moles (Condylura cristata) can sample odorants while malian evolution, olfaction became a less and less important sen- submerged using underwater sniffing (Catania, 2006). This ability sory modality, especially in primates, given its relatively reduced was thought impossible until these eulipotyphlans were filmed in “importance” as gauged from the relative volume of the olfactory slow-motion while foraging underwater. This remarkable under- bulb. This notion is disputed by our finding that the number of water olfactory ability likely stems from behavior specializations neurons in the olfactory bulb can be very large in eulipotyphlans, and the mechanics of the naris, rather than hypertrophied olfac- despite the relatively small size of the structure. tory structures per se. Thus, it is not inconsistent with lower Importantly, our data also dispute the notion that primates are densities of neurons in the olfactory bulb of star-nosed moles, not as reliant on olfaction, a notion that has been based on the which are clearly touch specialists (Catania and Kaas, 1997). In small relative size of the human olfactory bulb, compared to the any case, there is little doubt olfaction plays a special role for brain (Baron et al., 1983), and recently supported by the finding eulipotyphlans. that the loss of olfactory receptor genes coincides with the acqui- Our data thus indicate the occurrence of mosaic evolu- sition of full trichromatic vision in primates (Gilad et al., 2004). tion in the branching of eulipotyphlans from the common Those authors found that a twice higher percentage of olfactory Laurasiatherian ancestor, with the olfactory bulb and cerebel- receptor genes (32 against 16%) are pseudogenes in Old World lum scaling differently from the common ancestor while the monkeys and apes (which have full trichromatic vision) than in cerebral cortex and rest of brain retained the scaling rules also New World monkeys (which don’t have full trichromatic vision). found in afrotherians and glires. Nevertheless, even though our Although we were unable to examine the human olfactory bulb results suggest that the afrotherian/glires scaling rules applied in this study, the neuronal scaling rules that apply to primates to the olfactory bulb (and also cerebral cortex; see Neves et al., allow us to predict a total of 15–16 million neurons in the human 2014) of the ancestral eutherians, one should remain cautious olfactory bulb, given an average volume of 114 mm3 (which can when predicting the relative importance of olfaction in early be approximated to 114 mg; Baron et al., 1983). Surprisingly, this mammals from sheer volume of brain structures, given that number of neurons falls within the same range as the number of structure size does not predict number of neurons homoge- neurons we found in the largest eulipotyphlan olfactory bulbs. neously across all mammals (Herculano-Houzel, 2011). It will be Thus, while humans have been considered to be microsmatic

Frontiers in Neuroanatomy www.frontiersin.org April 2014 | Volume 8 | Article 23 | 10 Ribeiro et al. Numbers of neurons in the olfactory bulb

based on the relative (not absolute) size of their olfactory bulb, Catania, K. C., and Kaas, J. H. (1997). Somatosensory fovea in the star-nosed they still have as many olfactory bulb neurons as are found to sup- mole: behavioral use of the star in relation to innervation patterns and cor- port olfaction in animals that rely heavily on it for their behavior tical representation. J. Comp. Neurol. 387, 215–233. doi: 10.1002/(SICI)1096- 9861(19971020)387:2<215::AID-CNE4>3.0.CO;2-3 and survival. This similarity is in line with the observation that the Clark, D. A., Mitra, P. P., and Wang, S. S. (2001). Scalable architecture in mam- squirrel monkey has an unexpectedly high olfactory sensitivity malian brains. Nature 411, 189–193. doi: 10.1038/35075564 (Laska et al., 2000), and with the more recent finding that human Douady, C. J., Chatelier, C. I., Madsen, O., de Jong, W. W., Catzeflis, F., Springer, M. olfactory abilities are actually better than presumed (Porter et al., S., et al. (2002). Molecular phylogenetic evidence confirming the Eulipotyphla 2006). Gordon Shepherd challenged the belief that the behavioral concept and in support of hedgehogs as the sister group to shrews. Mol. Phylogen. Evol. 25, 200–209. doi: 10.1016/S1055-7903(02)00232-4 importance of the sense of olfaction diminished during human Finlay, B. L., and Darlington, R. B. (1995). Linked regularities in the devel- evolution, and defended the notion that human olfaction serves opment and evolution of mammalian brains. Science 268, 1578–1584. doi: food perception and seeking, as it is the main component of fla- 10.1126/science.7777856 vor (Shepherd, 2012). Thus, given the large absolute number of Finlay, B. L., Darlington, R. B., and Nicastro, N. (2001). Developmental neurons predicted to compose their olfactory bulb, compared to structure in brain evolution. Behav. Brain Sci. 24, 263–308. doi: 10.1017/S0140525X01003958 macrosmatic eulipotyphlans, humans (and other primates; Laska Gabi, M., Collins, C. E., Wong, P., Torres, L. B., Kaas, J. H., and Herculano-Houzel, et al., 2000) should no longer be considered microsmatic. Indeed, S. (2010). Cellular scaling rules for the brains of an extended number of primate the absence of olfaction changes human life in ways that transcend species. Brain Behav. Evol. 76, 32–44. doi: 10.1159/000319872 flavor perception, for instance depriving individuals of an olfac- Gilad, Y., Wiebe, V., Przeworski, M., Lancet, D., and Pääbo, S. (2004). Loss of olfactory receptor genes coincides with the acquisition of full trichromatic vision in tory perception of their surrounding environment, which goes primates. PLoS Biol 2:e5..doi: 10.1371/journal.pbio.0020005 beyond simple object identification (Tafalla, 2013). We suggest Herculano-Houzel, S. (2010). Coordinated scaling of cortical and cerebel- that a decrease in the relative size of the olfactory bulb is to be lar numbers of neurons. Front. Neuroanat. 4:12. doi: 10.3389/fnana. expected in primates given the faster addition of neurons to their 2010.00012 cerebral cortex than to other brain areas, but that does not imply Herculano-Houzel, S. (2011). Not all brains are made the same: new views on brain scalinginevolution.Brain Behav. Evol. 78, 22–36. doi: 10.1159/000327318 a diminished olfactory capacity. We propose that a distinction be Herculano-Houzel, S., Collins, C. E., Wong, P., and Kaas, J. H. (2007). Cellular scal- made between olfactory capabilities and relative reliance on olfac- ing rules for primate brains. Proc. Natl. Acad. Sci. U.S.A. 104, 3562–3567. doi: tion. Eulipotyphlans have both large olfactory capabilities and 10.1073/pnas.0611396104 rely heavily on them; primates still have strong olfactory capabil- Herculano-Houzel, S., and Lent, R. (2005). Isotropic fractionator: a simple, rapid ities, but rely more heavily on vision. These different strategies method for the quantification of total cell and neuron numbers in the brain. J. Neurosci. 25, 2518–2521. doi: 10.1523/JNEUROSCI.4526-04.2005 illustrate the principle of mosaic evolution in generating brain Herculano-Houzel, S., Mota, B., and Lent, R. (2006). Cellular scaling rules diversity: while expanding the cerebral cortex by adding neurons for rodent brains. Proc. Natl. Acad. Sci. U.S.A. 103, 12138–12143. doi: to it is expected to bring cognitive advantages, apparently eulipo- 10.1073/pnas.0604911103 typhlans do well enough with more neurons in the olfactory bulb Herculano-Houzel, S., Ribeiro, P., Campos, L., da Silva, A. V., Torres, L. B., than in the cerebral cortex. Catania, K. C., et al. (2011). Updated neuronal scaling rules for the brains of Glires (rodents/lagomorphs). Brain Behav. Evol. 78, 302–314. doi: 10.1159/ Finally, we show that the absolute expansion of the cerebral 000330825 cortex does not necessarily imply a relative expansion of the Jacobs, M. S., Morgane, P. J., and McFarland, W. L. (1971). The anatomy cerebral cortex in terms of the percentage of brain neurons con- of the brain of the bottlenose doplhin (Tursiops truncatus). Rhinic lobe centrated in the cerebral cortex. This finding adds to the previous (Rhinencephalon). I. The paleocortex. J. Comp. Neurol. 141, 205–272. doi: demonstration that, despite its relative expansion over the cere- 10.1002/cne.901410205 Jolicoeur, P., Pirlot, P., Baron, G., and Stephan, H. (1984). Brain structure and cor- bellum across species, the cerebral cortex actually gains neurons at relation patterns in Insectivora, Chiroptera, and Primates. Syst. Zool. 33, 14–29. a similar rate as the cerebellum across species (Herculano-Houzel, doi: 10.2307/2413129 2010). Thus, although cortical expansion certainly is a prominent Laska, M., Seibt, A., and Weber, A. (2000). ‘Microsmatic’ primates revisited: feature of brain evolution, it should no longer be equated with olfactory sensitivity in the squirrel monkey. Chem. Senses 25, 47–53. doi: 10.1093/chemse/25.1.47 thenotionofacorticaltakeoverattheexpenseofallotherbrain Lois, C., García-Verdugo, J. M., and Alvarez-Buylla, A. (1996). Chain migration of structures. neuronal precursors. Science 271, 978–981. doi: 10.1126/science.271.5251.978 Madsen, O., Scally, M., Douady, C. J., Kao, D. J., DeBry, R. W., Adkins, R., et al. REFERENCES (2001). Parallel adaptive radiations in two major clades of placental mammals. Bahney, J., and von Bartheld, C. S. (2014). Validation of the isotropic fractionator: Nature 409, 610–614. doi: 10.1038/35054544 comparison with unbiased stereology and DNA extraction for quantification of Mullen, R. J., Buck, C. R., and Smith, A. M. (1992). NeuN, a neuronal specific glial cells. J. Neurosci. Meth. 222, 165–174. doi: 10.1016/j.jneumeth.2013.11.002 nuclear protein in vertebrates. Development 116, 201–211. Baron, G., Frahm, H. D., Bhatnagar, K. P., and Stephan, H. (1983). Comparison of Murphy, W. J., Eizirik E, Johnson, W. E., Ping Zhang, Y., Ryder, O. A., and O’Brien, brain structure volumes in Insectivora and Primates. III. Main olfactory bulb S. J. (2001). Molecular phylogenetics and the origins of placental mammals. (MOB). J. Hirnforsch. 24, 551–568. Nature 409, 614–618. doi: 10.1038/35054550 Barton, R. A., and Harvey, P. H. (2000). Mosaic evolution of brain structure in Neves, K., Ferreira Meireles, F., Tovar-Moll, F., Gravett, N., Bennett, N. C., Kaswera, mammals. Nature 405, 1055–1058. doi: 10.1038/35016580 C., et al. (2014). Cellular scaling rules for the brain of afrotherians. Front. Butler, A. B., and Hodos, W. (2005). Comparative Vertebrate Neuroanatomy: Neuroanat. 8:5. doi: 10.3389/fnana.2014.00005 Evolution and Adaptation. New Jersey, NJ: Wiley-Interscience. doi: Parrish-Aungst, S., Shipley, M. T., Erdelyi, F., Szabo, G., and Puche, A. C. (2007). 10.1002/0471733849 Quantitative analysis of neuronal diversity in the mouse olfactory bulb. J. Comp. Catania, K. C. (2006). Olfaction: underwater ‘sniffing’ by semi-aquatic mammals. Neurol. 501, 825–836. doi: 10.1002/cne.21205 Nature 444, 1024–1025. doi: 10.1038/4441024a Porter, J., Craven, B., Khan, R. M., Chang, S.-J, Kang, I., Judkewitz, B., et al. Catania, K. C. (2013). Stereo and serial sniffing guide navigation to an odour source (2006). Mechanisms of scent-tracking in humans. Nat. Neurosci. 10, 27–29. doi: in a mammal. Nat. Commun. 4, 1441. doi: 10.1038/ncomms2444 10.1038/nn1819

Frontiers in Neuroanatomy www.frontiersin.org April 2014 | Volume 8 | Article 23 | 11 Ribeiro et al. Numbers of neurons in the olfactory bulb

Reep, R. L., Finlay, B. L., and Darlington, R. B. (2007). The limbic system Waddell, P. J., Cao, Y., Hauf, J., and Hasegawa, M. (1999). Using novel phylogenetic in mammalian brain evolution. Brain Behav. Evol. 70, 57–70. doi: 10.1159/ methods to evaluate mammalian mtDNA, including amino acid-invariant sites- 000101491 Logdet plus site stripping to detect internal conflicts in the data, with special Rowe, T. B., Macrini, T. E., and Luo, Z. X. (2011). Fossil evidence on origin of the reference to the positions of hedgehog, armadillo and elephant. Syst. Biol. 48, mammalian brain. Science 332, 955–957. doi: 10.1126/science.1203117 31–53. doi: 10.1080/106351599260427 Sarko, D. K., Catania, K. C., Leitch, D. B., Kaas, J. H., and Herculano-Houzel, S. Waddell, P. J., and Shelley, S. (2003). Evaluating placental inter-ordinal phylo- (2009). Cellular scaling rules of insectivore brains. Front. Neuroanat. 3:8. doi: genies with novel sequences including RAG1, gamma-fibrinogen, ND6, and 10.3389/neuro.05.008.2009 mtRNA, plus MCMC-driven nucleotide, amino acid, and codon models. Mol. Shepherd, G. M. (2012). Neurogastronomy: How the Brain Creates Flavor and why it Phylogenet. Evol. 28, 197–224. doi: 10.1016/S1055-7903(03)00115-5 Matters. New York, NY: Columbia University Press. Simpson, G. G. (1945). The principles of classification and a classification of Conflict of Interest Statement: The authors declare that the research was con- mammals. Bull. Am. Mus. Nat. Hist. 85, 1–350. ducted in the absence of any commercial or financial relationships that could be Stanhope, M. J., Waddell, V. G., Madsen O, de Jong, W. W., Hedges, S. construed as a potential conflict of interest. B., Cleven, G. C., et al. (1998). Molecular evidence for multipole origins of Insectivora and for a new order of endemic African insectivore Received: 30 January 2014; accepted: 24 March 2014; published online: 11 April 2014. mammals. Proc. Natl. Acad. Sci. U.S.A. 95, 9967–9972. doi: 10.1073/pnas. Citation: Ribeiro PFM, Manger PR, Catania KC, Kaas JH and Herculano-Houzel S 95.17.9967 (2014) Greater addition of neurons to the olfactory bulb than to the cerebral cortex of Stephan, H., and Andy, O. J. (1969). Quantitative comparative anatomy of primates: eulipotyphlans but not rodents, afrotherians or primates. Front. Neuroanat. 8:23. doi: an attempt at a phylogenetic interpretation. Ann. N.Y. Acad. Sci. 167, 370–386. 10.3389/fnana.2014.00023 doi: 10.1111/j.1749-6632.1969.tb20457.x This article was submitted to the journal Frontiers in Neuroanatomy. Sultan, F. (2002). Analysis of mammalian brain architecture. Nature 415, 133–134. Copyright © 2014 Ribeiro, Manger, Catania, Kaas and Herculano-Houzel. This is an doi: 10.1038/415133b open-access article distributed under the terms of the Creative Commons Attribution Tafalla, M. (2013). A world without the olfactory dimension. Anat. Rec. 296, License (CC BY). The use, distribution or reproduction in other forums is permitted, 1287–1296. doi: 10.1002/ar.22734 provided the original author(s) or licensor are credited and that the original publica- Turner, W. (1891). The convolutions of the brain: a study in comparative anatomy. tion in this journal is cited, in accordance with accepted academic practice. No use, J Anat Physiol 25, 105–153. distribution or reproduction is permitted which does not comply with these terms.

Frontiers in Neuroanatomy www.frontiersin.org April 2014 | Volume 8 | Article 23 | 12